CN117171693B - Cutting abnormality detection method in woodworking polishing process - Google Patents
Cutting abnormality detection method in woodworking polishing process Download PDFInfo
- Publication number
- CN117171693B CN117171693B CN202311412173.XA CN202311412173A CN117171693B CN 117171693 B CN117171693 B CN 117171693B CN 202311412173 A CN202311412173 A CN 202311412173A CN 117171693 B CN117171693 B CN 117171693B
- Authority
- CN
- China
- Prior art keywords
- cluster
- data
- class
- value
- center
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000005520 cutting process Methods 0.000 title claims abstract description 67
- 238000001514 detection method Methods 0.000 title claims abstract description 19
- 230000005856 abnormality Effects 0.000 title claims abstract description 16
- 238000007517 polishing process Methods 0.000 title claims abstract description 12
- 230000002159 abnormal effect Effects 0.000 claims abstract description 43
- 238000000034 method Methods 0.000 claims description 29
- 230000008569 process Effects 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000010586 diagram Methods 0.000 claims description 3
- 238000005498 polishing Methods 0.000 claims 3
- 230000000694 effects Effects 0.000 description 5
- 238000003064 k means clustering Methods 0.000 description 4
- 230000009466 transformation Effects 0.000 description 3
- 230000002776 aggregation Effects 0.000 description 2
- 238000004220 aggregation Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 239000012467 final product Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of data processing, in particular to a cutting abnormality detection method in the woodworking polishing process, which comprises the following steps: converting dynamic data in a cutting process into data points in a parameter space, acquiring class clusters in the parameter space under each k value, acquiring class clusters in an original space under each k value according to the class clusters in the parameter space under each k value, calculating the suitability of each k value according to the data point distribution condition in all class clusters in the original space under each k value, acquiring an optimal k value according to the suitability, acquiring an optimal clustering result of the dynamic data in the cutting process according to the optimal k value, and acquiring abnormal data points in the dynamic data in the cutting process according to the optimal clustering result to realize abnormal cutting detection. The invention obtains the class cluster with elliptical distribution state in the original space, separates the abnormal dynamic data from the normal dynamic data, improves the identification accuracy of the abnormal dynamic data, and further improves the accuracy of cutting abnormal detection.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a cutting abnormality detection method in the woodworking polishing process.
Background
Abnormal cutting in the woodworking polishing process can cause uneven, serrated or other defects on the surface of the woodworking, and the appearance and quality of the product are affected. By detecting the abnormality in the cutting process, the problems are found and repaired in time, and the final product is ensured to meet the requirements.
In the prior art, dynamic data in the cutting process is clustered, abnormal data is identified according to a clustering result, and then cutting abnormality detection is realized.
The K-means clustering algorithm is a conventional clustering algorithm and is suitable for clustering normal data with a round distribution state; however, the normal dynamic data has local linear correlation in the cutting process, so that the normal dynamic data distribution state is elliptical; therefore, the abnormal data in the dynamic data is identified by using the K-means clustering algorithm, and partial abnormal data can be incorrectly identified as normal data, so that the abnormal detection cannot be cut and is inaccurate.
Disclosure of Invention
The invention provides a cutting abnormality detection method in the woodworking polishing process, which aims to solve the existing problems, and comprises the following steps:
collecting dynamic data in the cutting process, and converting the dynamic data into data points in a parameter space, wherein each data point in the parameter space has a corresponding polar angle, polar diameter and voting rate;
obtaining class clusters in a parameter space under each k value comprises the following steps: setting a class cluster center according to the voting rate and the k value, calculating the characteristic similarity of each data point and each class cluster center in the parameter space according to the difference of the polar angle between the data point and the class cluster center and the difference of the polar diameter, and obtaining the class cluster in the parameter space according to the difference of the voting rate between each data point and each class cluster center and the characteristic similarity;
obtaining class clusters in the original space under each k value according to the class clusters in the parameter space under each k value, and calculating the suitability degree of each k value according to the data point distribution condition of all the class clusters in the original space under each k value;
and obtaining an optimal k value according to the suitability degree, obtaining an optimal clustering result of the dynamic data in the cutting process according to the optimal k value, and obtaining abnormal data points in the dynamic data in the cutting process according to the optimal clustering result to realize abnormal cutting detection.
Preferably, the setting the cluster center according to the voting rate and the k value includes the following specific steps:
taking the data point with the largest voting rate of all data points corresponding to each target polar angle as a cluster center, and obtaining k cluster centers, wherein the target polar angle is k polar angles with the largest voting rate, and the voting rate of each polar angle is the sum of the voting rates of all the data points corresponding to each polar angle in a parameter space.
Preferably, the calculation formula of the feature similarity between each data point in the parameter space and the center of each class cluster is as follows:
in the method, in the process of the invention,characteristic similarity of data points and cluster center is represented, < +.>And->Respectively representing the polar angle and the polar diameter of the data point, < ->And->Respectively representing the polar angle and the polar diameter of the cluster-like center, +.>Represents an exponential function based on natural constants, < ->Error tolerance indicating polar angle, +.>The error allowable range of the polar diameter is shown.
Preferably, the step of obtaining the class clusters in the parameter space according to the difference of the voting rate of each data point and the center of each class cluster and the feature similarity comprises the following specific steps:
for any data point in the parameter space, calculating the feature similarity between the data point and the center of each class cluster, dividing the data point into the class cluster center with the maximum feature similarity, and if the voting rate of the data point is inWithin the range, taking the data point as the data point in the cluster with the largest characteristic similarity in the cluster center>The voting rate of the cluster center with the largest feature similarity is represented;
and by analogy, dividing each data point in the parameter space into the class cluster where the center of the corresponding class cluster is located, and taking the class cluster as the class cluster in the parameter space.
Preferably, the obtaining the class cluster in the original space under each k value according to the class cluster in the parameter space under each k value includes the following specific steps:
obtaining class clusters in an original space according to the class clusters in the parameter space, and voting all data points in the original space for all data points in each class cluster in the parameter space to form one class cluster in the original space;
and removing abnormal points in each class cluster in the original space by using the box diagram, and dividing the data points into class clusters closest to the geometric center of each class cluster by calculating the distance between the data points and the geometric center of each class cluster.
Preferably, the calculation formula of the suitability degree of each k value is as follows:
in the method, in the process of the invention,indicating the appropriateness of the k value, the k value is taken over [3,10]All integers in>Mean value representing distance from all data points in ith class cluster in k class clusters to datum line of the class cluster,/for each data point in ith class cluster>Representing the number of data points in the ith class cluster in the k class clusters,/for each class cluster>Representing the area of the smallest circumscribed rectangle of the ith cluster in the k clusters, +.>And->The method respectively represents a long side and a short side of the smallest circumscribed rectangle of the ith class cluster in the k class clusters, wherein the long side is the side with the longest length of the two sides of the smallest circumscribed rectangle, and the short side is the side with the shortest length of the two sides of the smallest circumscribed rectangle.
Preferably, the method for acquiring the datum line of the cluster comprises the following steps:
and obtaining the minimum circumscribed rectangle of each cluster, and making a straight line parallel to the long side of the minimum circumscribed rectangle by passing through the geometric center of the minimum circumscribed rectangle of each cluster as a datum line of the cluster.
Preferably, the converting the dynamic data into data points in the parameter space includes the following specific steps:
the dynamic data in the cutting process comprises two dimensions of power output data and vibration data, the power output data and the vibration data form an original space, and the power output data and the vibration data at each moment are converted into the original spaceConverting data points in an original space into data points in a parameter space by Hough transformation, wherein the parameter space is formed by polar anglesAnd polar diameter->Is composed of the components.
The technical scheme of the invention has the beneficial effects that: according to the characteristic that normal dynamic data in the cutting process has local linear correlation, the dynamic data in the cutting process is converted into Hough transformation for clustering, and class clusters in Hough space are obtained by limiting the difference between data points and the cluster centers in polar angles and polar diameters, so that the class clusters with elliptic distribution states in the original space are obtained according to the class clusters in Hough space, abnormal dynamic data and normal dynamic data are separated, the recognition accuracy of the abnormal dynamic data is improved, and the accuracy of cutting abnormality detection is further improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flow chart of steps of a method for detecting cutting abnormality in a woodworking polishing process according to the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following is a detailed description of specific implementation, structure, characteristics and effects of the method for detecting abnormal cutting in the woodworking polishing process according to the invention with reference to the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following specifically describes a specific scheme of the method for detecting abnormal cutting in the woodworking polishing process provided by the invention with reference to the accompanying drawings.
In the prior art, dynamic data in the cutting process is clustered, abnormal data is identified according to a clustering result, and then cutting abnormality detection is realized. Because the normal dynamic data in the cutting process has local linear correlation, the distribution state of the normal dynamic data in the cutting process in the original space is elliptical, and the K-means clustering algorithm is suitable for clustering the normal data with circular distribution state; because the overall distribution state of the partial abnormal dynamic data and the normal dynamic data in the cutting process is circular when the partial abnormal dynamic data and the normal dynamic data are mixed together, if the partial abnormal dynamic data and the normal dynamic data are clustered in an original space through a K-means clustering algorithm, the partial abnormal dynamic data and the normal dynamic data are divided into a class cluster, and further when the abnormal data in the dynamic data are identified, the partial abnormal dynamic data are incorrectly identified as the normal dynamic data, the identification result of the abnormal dynamic data is inaccurate, and the cutting abnormality detection result is inaccurate, so that the abnormal dynamic data in the cutting process cannot be accurately identified by clustering in the original space.
The normal dynamic data has local linear correlation in the cutting process, so that the distribution state of the normal dynamic data in the original space is composed of a plurality of parallel straight lines, and the corresponding distribution state in the Hough space is a plurality of high-voting data points with similar characteristics.
Referring to fig. 1, a flowchart of a method for detecting abnormal cutting in a woodworking polishing process according to an embodiment of the invention is shown, the method includes the following steps:
s001, collecting dynamic data in the cutting process.
It should be noted that, the dynamic data in the cutting process refers to vibration data and power output data: vibration data generated by woodworking cutting equipment in operation is collected by a vibration sensor, and the vibration data is often used for analyzing the stability and balance of the cutting equipment and detecting whether abnormal vibration exists; the power output data of the cutting device in operation is passed through a power sensor, and the power output data is used to evaluate the working efficiency and energy consumption of the cutting device and to detect whether an abnormal power output condition exists.
In this embodiment, the dynamic data in the cutting process includes two dimensions of power output data and vibration data, the power output data and the vibration data form an original space, and the power output data and the vibration data generated by the woodworking cutting equipment at each moment in one minute in operation are converted into data points in the original space.
S002, converting the data points in the original space into data points in the parameter space, setting the center of the class cluster according to the voting rate and the k value, calculating the feature similarity between each data point in the parameter space and each class cluster center, and obtaining the class clusters in the parameter space according to the difference of the voting rate between each data point and each class cluster center and the feature similarity.
It should be noted that, under normal conditions, the greater the power output data generated by the woodworking cutting device in operation, the greater the vibration data, the two are in a forward linear relationship, and the power output data are in different ranges, and the degrees of the forward linear relationship between the power output data and the vibration data are different.
In the present embodiment, data points in an original space are converted into data points in a parameter space by hough transform, the parameter space is formed by polar anglesAnd polar diameter->Each data point has a voting rate in addition to the corresponding polar angle and polar diameter.
It should be noted that, since the dynamic data in the cutting process has a local linear correlation, the distribution state of the normal dynamic data in the original space is composed of a plurality of parallel straight lines, and the parallel straight lines represent a plurality of data points with higher voting rate in the parameter space, and the data points correspond to polar anglesIdentical or similar, and the corresponding polar diameters of the data points +.>And the characteristic similarity between the data points and the cluster center is calculated according to the similarity of the polar angle and the polar diameter, so that clustering is realized.
In this embodiment, among all the data points corresponding to each target polar angle, the data point with the largest voting rate is taken as a cluster center, k cluster centers are obtained in total, the target polar angle is k polar angles with the largest voting rate, and the voting rate of each polar angle is the sum of the voting rates of all the data points corresponding to each polar angle in the parameter space.
The feature similarity between each data point in the parameter space and the center of each class cluster is calculated according to the following specific calculation formula:
in the method, in the process of the invention,characteristic similarity of data points and cluster center is represented, < +.>And->Respectively representing the polar angle and the polar diameter of the data point, < ->And->Respectively representing the polar angle and the polar diameter of the cluster-like center, +.>Represents an exponential function based on natural constants, < ->Error tolerance indicating polar angle, +.>Representing the allowable range of the error of the polar angle and the allowable range of the error of the polar diameter are used for limiting the difference of the polar angle and the polar diameter of the data points belonging to the same cluster respectively, and in other embodiments, the practitioner can set the allowable range of the error of the polar angle and the allowable range of the error of the polar diameter according to the actual implementation condition, for example +>，。
It should be noted that, in order to make the cluster obtained in the original space be composed of normal dynamic data with local linear correlation, this embodiment will have polar anglesThe same or similar and the polar diameter +.>Several data points distributed in a small range are gathered into one type, so that the difference of polar angles and polar diameters between the data points and the center of the cluster are utilizedThe difference is that the cluster center of the data point is judged; difference in polar angle of data point and cluster-like center +.>The smaller the feature similarity between the data point and the center of the cluster is, the larger the feature similarity between the data point and the center of the cluster is; difference of polar diameter of data point and cluster-like center +.>The smaller the feature similarity between the data point and the center of the cluster is, the larger the feature similarity between the data point and the center of the cluster is; meanwhile, as the polar angle of the data point and the center of the cluster is required to be the same or similar, in order to ensure that the requirement is met, the difference between the polar angle of the data point and the center of the cluster is +.>The allowable error range of the adjustment electrode diameter is +.>Difference in polar angle of data point and cluster-like center +.>The larger the error allowance range of the adjusted polar diameter is, the smaller the error allowance range of the adjusted polar diameter is, and the difference between the data point and the polar diameter of the cluster-like center is adjusted to be +>The larger the difference between the pole diameters of the data points and the center of the class cluster is, the smaller the feature similarity between the data points and the center of the class cluster is.
It should be noted that, in addition to the feature similarity between the data point and the center of each class cluster, the clustering needs to ensure that the voting rates of the data point and the center of the class cluster are equal or similar.
For any data point in the parameter space, calculating the feature similarity between the data point and the center of each class cluster, dividing the data point into the class cluster center with the maximum feature similarity, and if the voting rate of the data point is inWithin the range, taking the data point as the data point in the cluster with the largest characteristic similarity in the cluster center>And the voting rate of the cluster center with the largest feature similarity is represented.
And by analogy, dividing each data point in the parameter space into the class cluster where the center of the corresponding class cluster is located, and taking the class cluster as the class cluster in the parameter space.
S003, class clusters in the original space under each k value are obtained according to the class clusters in the parameter space under each k value, the suitability degree of each k value is calculated according to the data point distribution condition of all class clusters in the original space under each k value, the optimal k value is obtained according to the suitability degree, and the optimal clustering result of dynamic data in the cutting process is obtained according to the optimal k value.
In this embodiment, class clusters in the original space are obtained from class clusters in the parameter space, and data points in all original spaces, which vote for all data points in each class cluster in the parameter space, constitute one class cluster in the original space.
It should be noted that, since one data point in the parameter space represents one straight line in the original space, and one data point in the original space may be on multiple straight lines in the original space, one data point in the original space may vote for multiple data points in the parameter space, and thus one data point in the original space may be divided into multiple clusters in the original space, and thus, it is necessary to determine the cluster to which the data point in the original space voted for multiple data points in the parameter space belongs.
In this embodiment, abnormal points in each class cluster in the original space are removed by using a box diagram, and the data points are divided into class clusters closest to each other by calculating the distance between the data points and the geometric center of each class cluster.
It should be noted that, if the number k of the clusters is different, the clustering results of the data points in the original space are different, the corresponding detection results of the cutting abnormality are also different, and as the number k of the clusters is increased, the data points in the original space are divided more finely, the aggregation degree of each cluster is gradually increased, when the number k of the clusters reaches the number of the real clusters, the aggregation degree obtained by increasing the k value is suddenly reduced, and in order to improve the accuracy of the cutting abnormality detection, the accuracy of the clustering needs to be improved, therefore, the data points in the original space are divided into the clusters of different numbers through the number k of the clusters, and then the appropriate degree of the k value of each cluster is calculated.
In this embodiment, according to the distribution situation of data points of k class clusters in the original space, the suitability of each k value is calculated, and a specific calculation formula is as follows:
in the method, in the process of the invention,indicating the appropriateness of the k value, the k value is taken over [3,10]All integers in>Mean value representing distance from all data points in ith class cluster in k class clusters to datum line of the class cluster,/for each data point in ith class cluster>Representing the number of data points in the ith class cluster in the k class clusters,/for each class cluster>Representing the area of the smallest circumscribed rectangle of the ith cluster in the k clusters, +.>And->The method respectively represents a long side and a short side of the smallest circumscribed rectangle of the ith class cluster in the k class clusters, wherein the long side is the side with the longest length of the two sides of the smallest circumscribed rectangle, and the short side is the side with the shortest length of the two sides of the smallest circumscribed rectangle.
And obtaining the minimum circumscribed rectangle of each cluster, and making a straight line parallel to the long side of the minimum circumscribed rectangle by passing through the geometric center of the minimum circumscribed rectangle of each cluster as a datum line of the cluster.
It should be noted that, the datum line of the cluster is the position where the data points in the cluster should be distributed intensively, and since the dynamic data in the cutting process has local linear correlation, the normal dynamic data in the original space is distributed near the datum line of the cluster, and the smaller the distance from the data point in the cluster to the datum line of the cluster is, the more the normal dynamic data in the original space is distributed near the datum line of the cluster, and the better the clustering effect is;the density of data point distribution in the characteristic class cluster is larger, and the clustering effect is better; />The linear degree of the data point distribution in the class cluster is characterized, and the larger the value is, the larger the linear degree of the data point distribution in the class cluster is, and the better the clustering effect is.
And taking the k value with the greatest suitability as the optimal k value, and obtaining the optimal clustering result of the dynamic data in the cutting process according to the optimal k value.
S004, abnormal data points in dynamic data in the cutting process are obtained according to the optimal clustering result, and cutting abnormality detection is achieved.
And obtaining abnormal data points according to the optimal clustering result of the dynamic data in the cutting process, and taking the data points which do not belong to any cluster as abnormal data points to realize the cutting abnormal detection.
According to the characteristic that normal dynamic data in the cutting process has local linear correlation, the dynamic data in the cutting process is converted into Hough transformation for clustering, and class clusters in Hough space are obtained by limiting the difference between data points and the cluster centers in polar angles and polar diameters, so that the class clusters with elliptic distribution states in the original space are obtained according to the class clusters in Hough space, abnormal dynamic data and normal dynamic data are separated, the recognition accuracy of the abnormal dynamic data is improved, and the accuracy of cutting abnormality detection is further improved.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the invention, but any modifications, equivalent substitutions, improvements, etc. within the principles of the present invention should be included in the scope of the present invention.
Claims (5)
1. The method for detecting the cutting abnormality in the woodworking polishing process is characterized by comprising the following steps of:
collecting dynamic data in the cutting process, and converting the dynamic data into data points in a parameter space, wherein each data point in the parameter space has a corresponding polar angle, polar diameter and voting rate;
obtaining class clusters in a parameter space under each k value comprises the following steps: setting a cluster center according to the voting rate and a k value, wherein the k value refers to the number of clusters, calculating the characteristic similarity between each data point and each cluster center in a parameter space according to the difference of polar angles and the difference of polar diameters between the data points and the cluster center, and obtaining the clusters in the parameter space according to the difference of the voting rate between each data point and each cluster center and the characteristic similarity;
obtaining class clusters in the original space under each k value according to the class clusters in the parameter space under each k value, and calculating the suitability degree of each k value according to the data point distribution condition of all the class clusters in the original space under each k value;
obtaining an optimal k value according to the suitability degree, obtaining an optimal clustering result of dynamic data in the cutting process according to the optimal k value, and obtaining abnormal data points in the dynamic data in the cutting process according to the optimal clustering result to realize abnormal cutting detection;
the method for converting dynamic data into data points in a parameter space comprises the following specific steps:
the dynamic data in the cutting process comprises two dimensions of power output data and vibration data, wherein the power output data and the vibration data form an original space, the power output data and the vibration data at each moment are converted into data points in the original space, and the data points pass through the hallThe Fu transform converts data points in the original space into data points in a parameter space, which is defined by polar anglesAnd polar diameter->Is composed of;
the calculation formula of the suitability degree of each k value is as follows:
in the method, in the process of the invention,indicating the appropriateness of the k value, the k value is taken over [3,10]All integers in>Mean value representing distance from all data points in ith class cluster in k class clusters to datum line of the class cluster,/for each data point in ith class cluster>Representing the number of data points in the ith class cluster in the k class clusters,/for each class cluster>Representing the area of the smallest circumscribed rectangle of the ith cluster in the k clusters, +.>And->Respectively representing the long side and the short side of the smallest circumscribed rectangle of the ith class cluster in the k class clusters, wherein the long side is the side with the longest length of the two sides of the smallest circumscribed rectangle, and the short side is the side with the shortest length of the two sides of the smallest circumscribed rectangle;
the method for acquiring the datum line of the cluster comprises the following steps: and obtaining the minimum circumscribed rectangle of each cluster, and making a straight line parallel to the long side of the minimum circumscribed rectangle by passing through the geometric center of the minimum circumscribed rectangle of each cluster as a datum line of the cluster.
2. The method for detecting abnormal cutting in the woodworking polishing process according to claim 1, wherein the setting of the cluster center according to the voting rate and the k value comprises the following specific steps:
taking the data point with the largest voting rate of all data points corresponding to each target polar angle as a cluster center, and obtaining k cluster centers, wherein the target polar angle is k polar angles with the largest voting rate, and the voting rate of each polar angle is the sum of the voting rates of all the data points corresponding to each polar angle in a parameter space.
3. The method for detecting abnormal cutting in the process of woodworking polishing according to claim 1, wherein the calculation formula of the feature similarity between each data point in the parameter space and each cluster-like center is as follows:
in the method, in the process of the invention,characteristic similarity of data points and cluster center is represented, < +.>And->Respectively representing the polar angle and the polar diameter of the data point, < ->And->Respectively representing the polar angle and the polar diameter of the cluster-like center, +.>Represents an exponential function based on natural constants, < ->Error tolerance indicating polar angle, +.>The error allowable range of the polar diameter is shown.
4. The method for detecting abnormal cutting in the process of woodworking polishing according to claim 1, wherein the step of obtaining the class clusters in the parameter space according to the difference of the voting rate of each data point and the center of each class cluster and the feature similarity comprises the following specific steps:
for any data point in the parameter space, calculating the feature similarity between the data point and the center of each class cluster, dividing the data point into the class cluster center with the maximum feature similarity, and if the voting rate of the data point is inWithin the range, taking the data point as the data point in the cluster with the largest characteristic similarity in the cluster center>The voting rate of the cluster center with the largest feature similarity is represented;
and by analogy, dividing each data point in the parameter space into the class cluster where the center of the corresponding class cluster is located, and taking the class cluster as the class cluster in the parameter space.
5. The method for detecting abnormal cutting during woodworking polishing according to claim 1, wherein the step of obtaining the class cluster in the original space under each k value according to the class cluster in the parameter space under each k value comprises the following specific steps:
obtaining class clusters in an original space according to the class clusters in the parameter space, and voting all data points in the original space for all data points in each class cluster in the parameter space to form one class cluster in the original space;
and removing abnormal points in each class cluster in the original space by using the box diagram, and dividing the data points into class clusters closest to the geometric center of each class cluster by calculating the distance between the data points and the geometric center of each class cluster.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311412173.XA CN117171693B (en) | 2023-10-30 | 2023-10-30 | Cutting abnormality detection method in woodworking polishing process |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311412173.XA CN117171693B (en) | 2023-10-30 | 2023-10-30 | Cutting abnormality detection method in woodworking polishing process |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117171693A CN117171693A (en) | 2023-12-05 |
| CN117171693B true CN117171693B (en) | 2024-01-26 |
Family
ID=88937758
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311412173.XA Active CN117171693B (en) | 2023-10-30 | 2023-10-30 | Cutting abnormality detection method in woodworking polishing process |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117171693B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103531198A (en) * | 2013-11-01 | 2014-01-22 | 东南大学 | Speech emotion feature normalization method based on pseudo speaker clustering |
| CN103969069A (en) * | 2014-04-26 | 2014-08-06 | 常州大学 | Working condition monitoring and fault diagnosis method for mechanical equipment |
| CN109614948A (en) * | 2018-12-19 | 2019-04-12 | 北京锐安科技有限公司 | Detection method, device, equipment and the storage medium of abnormal behaviour |
| CN110517317A (en) * | 2019-08-28 | 2019-11-29 | 武汉西尔塔信息技术有限公司 | A kind of steel billet localization method and application based on the processing of machine vision camera image |
| CN114444382A (en) * | 2021-12-09 | 2022-05-06 | 中广核(北京)新能源科技有限公司 | Wind turbine generator gearbox fault diagnosis and analysis method based on machine learning algorithm |
| CN115270986A (en) * | 2022-08-08 | 2022-11-01 | 贵州电网有限责任公司电力科学研究院 | Data anomaly detection method and device and computer equipment |
| CN116071694A (en) * | 2023-03-07 | 2023-05-05 | 浙江华是科技股份有限公司 | Ship detection method, device and computer readable storage medium |
| CN116166960A (en) * | 2023-02-07 | 2023-05-26 | 河南大学 | Big data characteristic cleaning method and system for neural network training |
| CN116645367A (en) * | 2023-07-27 | 2023-08-25 | 山东昌啸商贸有限公司 | Steel plate cutting quality detection method for high-end manufacturing |
-
2023
- 2023-10-30 CN CN202311412173.XA patent/CN117171693B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103531198A (en) * | 2013-11-01 | 2014-01-22 | 东南大学 | Speech emotion feature normalization method based on pseudo speaker clustering |
| CN103969069A (en) * | 2014-04-26 | 2014-08-06 | 常州大学 | Working condition monitoring and fault diagnosis method for mechanical equipment |
| CN109614948A (en) * | 2018-12-19 | 2019-04-12 | 北京锐安科技有限公司 | Detection method, device, equipment and the storage medium of abnormal behaviour |
| CN110517317A (en) * | 2019-08-28 | 2019-11-29 | 武汉西尔塔信息技术有限公司 | A kind of steel billet localization method and application based on the processing of machine vision camera image |
| CN114444382A (en) * | 2021-12-09 | 2022-05-06 | 中广核(北京)新能源科技有限公司 | Wind turbine generator gearbox fault diagnosis and analysis method based on machine learning algorithm |
| CN115270986A (en) * | 2022-08-08 | 2022-11-01 | 贵州电网有限责任公司电力科学研究院 | Data anomaly detection method and device and computer equipment |
| CN116166960A (en) * | 2023-02-07 | 2023-05-26 | 河南大学 | Big data characteristic cleaning method and system for neural network training |
| CN116071694A (en) * | 2023-03-07 | 2023-05-05 | 浙江华是科技股份有限公司 | Ship detection method, device and computer readable storage medium |
| CN116645367A (en) * | 2023-07-27 | 2023-08-25 | 山东昌啸商贸有限公司 | Steel plate cutting quality detection method for high-end manufacturing |
Non-Patent Citations (7)
| Title |
|---|
| Conditional feature disentanglement learning for anomaly detection in machines operating under time-varying conditions;Haoxuan Zhou 等;《Mechanical Systems and Signal Processing》;第191卷;1-26 * |
| Unsupervised Anomaly Isolation for Condition Monitoring Systems;Emad Ali 等;《2018 2nd European Conference on Electrical Engineering and Computer Science (EECS)》;491-496 * |
| 基于改进K均值聚类的异常检测算法;左进 等;《计算机科学》;第43卷(第8期);258-261 * |
| 基于改进密度聚类与模式信息挖掘的异常轨迹识别方法;何明 等;《通信学报》;第38卷(第12期);21-33 * |
| 基于数学形态学的三维空间聚类;张永梅 等;《火力与指挥控制》(第10期);41-43+47 * |
| 基于深度学习的轨道扣件实时检测系统设计;饶鑫;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;第2022年卷(第5期);C033-127 * |
| 密度峰值快速聚类算法优化研究;王鹏飞 等;《计算机工程与科学》;第40卷(第8期);1503-1510 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117171693A (en) | 2023-12-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109579220B (en) | Air conditioning system fault detection method and device and electronic equipment | |
| CN110795690A (en) | Wind power plant operation abnormal data detection method | |
| CN110674120B (en) | Wind farm data cleaning method and device | |
| CN110348488B (en) | Modal identification method based on local density peak clustering | |
| CN114114039B (en) | Method and device for evaluating consistency of single battery cells of battery system | |
| CN115311262A (en) | Printed circuit board defect identification method | |
| CN113269791B (en) | Point cloud segmentation method based on edge judgment and region growth | |
| CN115359807A (en) | Noise online monitoring system for urban noise pollution | |
| CN113537321B (en) | Network flow anomaly detection method based on isolated forest and X mean value | |
| CN110852388A (en) | Improved SMOTE algorithm based on K-means | |
| CN116404186B (en) | Power lithium-manganese battery production system | |
| CN112257585A (en) | Breaker state monitoring method based on probability density | |
| CN109949324A (en) | Profile testing method based on the non-linear subunit response of non-classical receptive field | |
| CN115290316A (en) | Fault diagnosis method for eccentric rotary valve | |
| CN117171693B (en) | Cutting abnormality detection method in woodworking polishing process | |
| CN117170979B (en) | Energy consumption data processing method, system, equipment and medium for large-scale equipment | |
| WO2021258961A1 (en) | Network traffic classification method and system based on improved k-means algorithm | |
| CN114200245A (en) | Construction method of line loss abnormity identification model of power distribution network | |
| CN115660898B (en) | SVR-based classification type wind power short-term power prediction precision improving method and device | |
| CN114662624B (en) | Non-invasive load identification method and system based on retraining twin network | |
| CN111600051B (en) | Proton exchange membrane fuel cell fault diagnosis method based on image processing | |
| CN115129503A (en) | Equipment fault data cleaning method and system | |
| CN114710313A (en) | Equipment attack detection method, device and system for power grid | |
| CN111563543B (en) | Method and device for cleaning wind speed-power data of wind turbine generator | |
| CN111597549A (en) | Network security behavior identification method and system based on big data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |